La serie de la temperatura fue extraĆda de la base de datos del IDEAM, se tuvieron en cuenta los promedios diarios de las temperaturas en grados centigrados (Ā°C) registradas en BogotĆ” en las diferentes estaciones meteorolĆ³gicas que recolectaron informaciĆ³n de esos dĆas. La serie de tiempo cuenta con un total de 1826 registros, de los cuales 13 (0.7%) fueron imputados puesto que no se presentaba la informaciĆ³n necesaria. Esta imputaciĆ³n fue realizada a partir del mĆ©todo de vecinos mĆ”s cercanos (KNN), donde se tuvierĆ³n en cuenta 5 vecinos.
Para el modelar la serie de la temperatura se usara inicialmente un modelo SARIMA, seguido de redes recurrentes simples, GRU y finalmente GRU. Para seleccionar el mejor modelo se tomarƔ como criterio el error cuadrƔtico medio.
import numpy as np
import pandas as pd
import plotly.express as px
import matplotlib.pyplot as plt
import sklearn as sk
from sklearn import impute
from sklearn import preprocessing
import sklearn.externals
import joblib
from sklearn.model_selection import TimeSeriesSplit
from sklearn.impute import KNNImputer
import sklearn.preprocessing
from keras.models import Sequential
from tensorflow.keras.callbacks import EarlyStopping
import time
import sklearn.externals
import joblib
import plotly.graph_objects as go
from sklearn import metrics
import statsmodels.api as sm
import statsmodels.tsa.stattools as ts
from tensorflow.keras.models import Model
from tensorflow.keras.optimizers import Adam
from statsmodels.tsa.statespace.sarimax import SARIMAX
import tensorflow as tf
import tensorflow.keras.layers as L
import tensorflow.keras.models as M
import tensorflow.keras.backend as K
%matplotlib inline
import plotly.io as pio
pio.renderers.default='notebook'
from google.colab import drive
drive.mount('/content/drive')
Mounted at /content/drive
df = pd.read_csv("/content/drive/Shareddrives/Mineria /Temperatura1.csv", sep=';', header=0, decimal = ',')
Fecha = pd.date_range(start='2017-01-01', end='2021-12-31', freq='D')
df['Fecha'] = Fecha
df = df.set_index('Fecha')
print(df[pd.isnull(df.ValorObservado)])
print('En total hay' ,
str(df['ValorObservado'].isnull().sum()) ,
'valores sin informaciĆ³n')
print('Correspondientes al {:.3f}% del total'
.format(df['ValorObservado'].isnull().sum()*100/len(df)))
ValorObservado Fecha 2017-08-12 NaN 2017-12-24 NaN 2019-09-15 NaN 2019-09-16 NaN 2019-09-17 NaN 2020-11-12 NaN 2021-01-05 NaN 2021-01-06 NaN 2021-01-07 NaN 2021-01-08 NaN 2021-08-18 NaN 2021-08-20 NaN 2021-12-05 NaN En total hay 13 valores sin informaciĆ³n Correspondientes al 0.712% del total
La serie presenta valores faltantes, por lo tantol se imputaran usando el mĆ©todo de vecinos mĆ”s cercanos (KNN), como se muestra a continuaciĆ³n.
#ImputaciĆ³n de Valores usando el vecino mĆ”s cercano
imput = KNNImputer(n_neighbors=5, weights="uniform")
# Ajustamos el modelo e imputamos los missing values
imput.fit(df[['ValorObservado']])
df['ValorObservado'] = imput.transform(df[['ValorObservado']]).ravel()
print()
print("Valores pƩrdidos en ValorObservado: " ,
str(df['ValorObservado'].isnull().sum()))
Valores pƩrdidos en ValorObservado: 0
fig = px.line(df, x=df.index, y="ValorObservado")
fig.update_xaxes(title_text="Fecha")
fig.show()
Para el respectivo anƔlisis se tomarƔn el 80% de los datos para entrenamiento y validacion, el 20% restantes para prueba, dichos valores corresponden a 1460 y 365 respectivamente.
from sklearn.preprocessing import MinMaxScaler
# crea el objeto scaler y escala los datos
scaler = MinMaxScaler(feature_range=(0, 1))
scaled_data = scaler.fit_transform(df.values)
#
df_norm = pd.DataFrame(scaled_data,index=df.index, columns=['ValorObservadoNormalizado'])
train_size = int(len(df_norm) * 0.8)
test_size = len(df_norm) - train_size
train, test = df_norm.iloc[0:train_size], df_norm.iloc[train_size:len(df_norm)]
len_train = len(train)
len_test = len(test)
print(len_train, len_test)
1460 366
def create_dataset(X, y, time_steps=1):
# crea dos listas vacias para depositar los datos
Xs, ys = [], []
# el primer lote de datos empieza en la primera observaciĆ³n
# y toma time_steps datos.
# Comienza a avanzar hacia adelante.
for i in range(len(X) - time_steps):
v = X.iloc[i:(i + time_steps)].values
Xs.append(v)
ys.append(y.iloc[i + time_steps])
return np.array(Xs), np.array(ys)
time_steps = 50
# reshape to [samples, time_steps, n_features]
X_train, y_train = create_dataset(train, train, time_steps)
X_test, y_test = create_dataset(test, test, time_steps)
print("X_train.shape = ", X_train.shape)
print("y_train.shape = ", y_train.shape)
print("X_test.shape = ", X_test.shape)
print("y_test.shape = ", y_test.shape)
X_train.shape = (1410, 50, 1) y_train.shape = (1410, 1) X_test.shape = (316, 50, 1) y_test.shape = (316, 1)
fig = px.line(df_norm, x=df.index, y='ValorObservadoNormalizado')
fig.update_xaxes(title_text="Fecha")
fig.update_yaxes(title_text="ValorObservadoNormalizado")
fig.show()
# shapes
inputs_shape = (X_train.shape[1], X_train.shape[2])
GRU_output = 60
# layers
inputs = L.Input(inputs_shape)
gru = L.GRU(units=GRU_output, name='GRU')(inputs)
outputs_GRU = L.Dense(1)(gru)
GRU_model = Model(inputs=inputs, outputs=outputs_GRU, name='series_GRU_model')
# Compiling the RNN
GRU_model.compile(loss="mean_squared_error", optimizer=Adam(0.001))
# Fitting to the training set
start = time.time()
GRU = GRU_model.fit(
X_train,
y_train,
epochs=50,
batch_size=16,
validation_split=0.1,
verbose=1,
shuffle=False
)
print("compilation time : ", time.time() - start)
Epoch 1/50 80/80 [==============================] - 5s 28ms/step - loss: 0.0182 - val_loss: 0.0038 Epoch 2/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0068 - val_loss: 0.0037 Epoch 3/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0064 - val_loss: 0.0034 Epoch 4/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0061 - val_loss: 0.0032 Epoch 5/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0060 - val_loss: 0.0032 Epoch 6/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0060 - val_loss: 0.0031 Epoch 7/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 8/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 9/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 10/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 11/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 12/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 13/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 14/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 15/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 16/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 17/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 18/50 80/80 [==============================] - 3s 33ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 19/50 80/80 [==============================] - 3s 31ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 20/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 21/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 22/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 23/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 24/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 25/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 26/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 27/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 28/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 29/50 80/80 [==============================] - 2s 23ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 30/50 80/80 [==============================] - 2s 26ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 31/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 32/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 33/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 34/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 35/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 36/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 37/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 38/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 39/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 40/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0030 Epoch 41/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0030 Epoch 42/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0030 Epoch 43/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0030 Epoch 44/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0030 Epoch 45/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0030 Epoch 46/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0058 - val_loss: 0.0030 Epoch 47/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0058 - val_loss: 0.0030 Epoch 48/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0057 - val_loss: 0.0030 Epoch 49/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0057 - val_loss: 0.0030 Epoch 50/50 80/80 [==============================] - 2s 21ms/step - loss: 0.0057 - val_loss: 0.0030 compilation time : 92.24258947372437
joblib.dump(GRU_model,'/content/drive/Shareddrives/Mineria /GRU')
WARNING:absl:Found untraced functions such as gru_cell_layer_call_fn, gru_cell_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.
['/content/drive/Shareddrives/Mineria /GRU']
GRU_model.summary()
Model: "series_GRU_model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_1 (InputLayer) [(None, 50, 1)] 0
GRU (GRU) (None, 60) 11340
dense (Dense) (None, 1) 61
=================================================================
Total params: 11,401
Trainable params: 11,401
Non-trainable params: 0
_________________________________________________________________
GRU_losses = pd.DataFrame(GRU.history)
fig = px.line(GRU_losses, x=GRU_losses.index, y=["loss", "val_loss"])
fig.update_xaxes(title_text="Epoch")
fig.update_yaxes(title_text="Loss")
fig.show()
GRU_Predict = GRU_model.predict(X_test)
GRU_Predict = scaler.inverse_transform(GRU_Predict)
10/10 [==============================] - 0s 7ms/step
# plot y_train, y_test, and testPredict using plotly
seq_len=50
fig = go.Figure()
fig.add_trace(
go.Scatter(
x=df.index[seq_len : len(y_train) + seq_len],
y=scaler.inverse_transform(y_train).ravel(),
mode="lines",
name="Entrenamiento",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train) + seq_len :],
y=scaler.inverse_transform(y_test).ravel(),
mode="lines",
name="Prueba",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train) + seq_len :],
y=GRU_Predict.ravel(),
mode="lines",
name="PredicciĆ³n",
)
)
fig.update_xaxes(title_text="Fecha")
fig.update_yaxes(title_text="ValorObservado")
fig.show()
def QuantileLoss(perc, delta=1e-4):
perc = np.array(perc).reshape(-1)
perc.sort()
perc = perc.reshape(1, -1)
def _qloss(y, pred):
I = tf.cast(y <= pred, tf.float32)
d = K.abs(y - pred)
correction = I * (1 - perc) + (1 - I) * perc
# huber loss
huber_loss = K.sum(correction * tf.where(d <= delta, 0.5 * d ** 2 / delta, d - 0.5 * delta), -1)
# order loss
q_order_loss = K.sum(K.maximum(0.0, pred[:, :-1] - pred[:, 1:] + 1e-6), -1)
return huber_loss + q_order_loss
return _qloss
# quantiles
perc_points = [0.025, 0.975]
# shapes
inputs_shape = (X_train.shape[1], X_train.shape[2])
GRU_output = 60
# layers
inputs = L.Input(inputs_shape)
qgru = L.GRU(units=GRU_output, name='GRU')(inputs)
qoutputs_GRU = L.Dense(2)(qgru)
qGRU_model = Model(inputs=inputs, outputs=qoutputs_GRU, name='series_GRU_model')
# Compiling the RNN
qGRU_model.compile(Adam(0.001), loss=QuantileLoss(perc_points))
# Fitting to the training set
start = time.time()
qGRU = qGRU_model.fit(
X_train,
y_train,
epochs=50,
batch_size=16,
validation_split=0.1,
verbose=1,
shuffle=False
)
print("compilation time : ", time.time() - start)
Epoch 1/50 80/80 [==============================] - 5s 28ms/step - loss: 0.1017 - val_loss: 0.0134 Epoch 2/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0137 - val_loss: 0.0083 Epoch 3/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0125 - val_loss: 0.0085 Epoch 4/50 80/80 [==============================] - 2s 27ms/step - loss: 0.0123 - val_loss: 0.0086 Epoch 5/50 80/80 [==============================] - 3s 33ms/step - loss: 0.0120 - val_loss: 0.0082 Epoch 6/50 80/80 [==============================] - 2s 23ms/step - loss: 0.0117 - val_loss: 0.0081 Epoch 7/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0115 - val_loss: 0.0076 Epoch 8/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0112 - val_loss: 0.0078 Epoch 9/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0111 - val_loss: 0.0077 Epoch 10/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0111 - val_loss: 0.0076 Epoch 11/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0109 - val_loss: 0.0075 Epoch 12/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0107 - val_loss: 0.0073 Epoch 13/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0108 - val_loss: 0.0072 Epoch 14/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0106 - val_loss: 0.0071 Epoch 15/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0106 - val_loss: 0.0073 Epoch 16/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0105 - val_loss: 0.0072 Epoch 17/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0106 - val_loss: 0.0072 Epoch 18/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0105 - val_loss: 0.0071 Epoch 19/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0105 - val_loss: 0.0072 Epoch 20/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0107 - val_loss: 0.0072 Epoch 21/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0105 - val_loss: 0.0072 Epoch 22/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0105 - val_loss: 0.0069 Epoch 23/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0106 - val_loss: 0.0071 Epoch 24/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0069 Epoch 25/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0105 - val_loss: 0.0069 Epoch 26/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0069 Epoch 27/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0105 - val_loss: 0.0072 Epoch 28/50 80/80 [==============================] - 2s 23ms/step - loss: 0.0104 - val_loss: 0.0072 Epoch 29/50 80/80 [==============================] - 3s 33ms/step - loss: 0.0104 - val_loss: 0.0070 Epoch 30/50 80/80 [==============================] - 2s 27ms/step - loss: 0.0104 - val_loss: 0.0070 Epoch 31/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0069 Epoch 32/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0067 Epoch 33/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0105 - val_loss: 0.0069 Epoch 34/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0067 Epoch 35/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0069 Epoch 36/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0067 Epoch 37/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0067 Epoch 38/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0069 Epoch 39/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0069 Epoch 40/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0068 Epoch 41/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0070 Epoch 42/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0070 Epoch 43/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0068 Epoch 44/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0068 Epoch 45/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0104 - val_loss: 0.0069 Epoch 46/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0069 Epoch 47/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0069 Epoch 48/50 80/80 [==============================] - 2s 23ms/step - loss: 0.0103 - val_loss: 0.0069 Epoch 49/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0069 Epoch 50/50 80/80 [==============================] - 2s 22ms/step - loss: 0.0103 - val_loss: 0.0068 compilation time : 93.8435800075531
qGRU_model.summary()
Model: "series_GRU_model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_2 (InputLayer) [(None, 50, 1)] 0
GRU (GRU) (None, 60) 11340
dense_1 (Dense) (None, 2) 122
=================================================================
Total params: 11,462
Trainable params: 11,462
Non-trainable params: 0
_________________________________________________________________
qGRU_losses = pd.DataFrame(qGRU.history)
fig = px.line(qGRU_losses, x=qGRU_losses.index, y=["loss", "val_loss"])
fig.update_xaxes(title_text="Epoch")
fig.update_yaxes(title_text="Loss")
fig.show()
qGRU_Predict = qGRU_model.predict(X_test)
qGRU_Predict = scaler.inverse_transform(qGRU_Predict)
10/10 [==============================] - 0s 7ms/step
# plot y_train, y_test, and testPredict using plotly
fig = go.Figure()
fig.add_trace(
go.Scatter(
x=df.index[seq_len : len(y_train) + seq_len],
y=scaler.inverse_transform(y_train).ravel(),
mode="lines",
name="Entrenamiento",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train) + seq_len :],
y=GRU_Predict.ravel(),
mode="lines",
name="PredicciĆ³n",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train) + seq_len :],
y=qGRU_Predict[:,0] ,
mode="lines",
name="0.025",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train) + seq_len :],
y=qGRU_Predict[:,1] ,
mode="lines",
name="0.975",
)
)
fig.update_xaxes(title_text="Fecha")
fig.update_yaxes(title_text="ValorObservado")
fig.show()
from sklearn.metrics import mean_squared_error
GRU_Score1 = mean_squared_error(scaler.inverse_transform(y_test), GRU_Predict)**.5
print('Test Score: %.2f RMSE' % (GRU_Score1))
Test Score: 0.84 RMSE
def create_dataset(X, y, time_steps=1):
# crea dos listas vacias para depositar los datos
Xs, ys = [], []
# el primer lote de datos empieza en la primera observaciĆ³n
# y toma time_steps datos.
# Comienza a avanzar hacia adelante.
for i in range(len(X) - time_steps-5):
v = X.iloc[i:(i + time_steps)].values
Xs.append(v)
ys.append(y.iloc[i + time_steps+5])
return np.array(Xs), np.array(ys)
time_steps = 50
# reshape to [samples, time_steps, n_features]
X_train_5p, y_train_5p = create_dataset(train, train, time_steps)
X_test_5p, y_test_5p = create_dataset(test, test, time_steps)
print("X_train.shape = ", X_train_5p.shape)
print("y_train.shape = ", y_train_5p.shape)
print("X_test.shape = ", X_test_5p.shape)
print("y_test.shape = ", y_test_5p.shape)
X_train.shape = (1405, 50, 1) y_train.shape = (1405, 1) X_test.shape = (311, 50, 1) y_test.shape = (311, 1)
# shapes
inputs_shape = (X_train_5p.shape[1], X_train_5p.shape[2])
GRU_output_5p = 60
# layers
inputs_5p = L.Input(inputs_shape)
gru_5p = L.GRU(units=GRU_output_5p, name='GRU')(inputs_5p)
outputs_GRU_5p = L.Dense(1)(gru_5p)
GRU_model_5p = Model(inputs=inputs_5p, outputs=outputs_GRU_5p, name='series_GRU_model')
# Compiling the RNN
GRU_model_5p.compile(optimizer=Adam(0.001),loss="mean_squared_error")
# Fitting to the training set
start = time.time()
GRU_5p = GRU_model_5p.fit(
X_train_5p,
y_train_5p,
epochs=50,
batch_size=16,
validation_split=0.1,
verbose=1,
shuffle=False
)
print("compilation time : ", time.time() - start)
Epoch 1/50 79/79 [==============================] - 3s 22ms/step - loss: 0.0286 - val_loss: 0.0052 Epoch 2/50 79/79 [==============================] - 2s 22ms/step - loss: 0.0123 - val_loss: 0.0055 Epoch 3/50 79/79 [==============================] - 2s 22ms/step - loss: 0.0122 - val_loss: 0.0055 Epoch 4/50 79/79 [==============================] - 1s 17ms/step - loss: 0.0122 - val_loss: 0.0055 Epoch 5/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0122 - val_loss: 0.0055 Epoch 6/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0121 - val_loss: 0.0055 Epoch 7/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0121 - val_loss: 0.0054 Epoch 8/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0121 - val_loss: 0.0054 Epoch 9/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0121 - val_loss: 0.0054 Epoch 10/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0120 - val_loss: 0.0054 Epoch 11/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0120 - val_loss: 0.0054 Epoch 12/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0120 - val_loss: 0.0054 Epoch 13/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0120 - val_loss: 0.0054 Epoch 14/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0119 - val_loss: 0.0053 Epoch 15/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0119 - val_loss: 0.0053 Epoch 16/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0119 - val_loss: 0.0053 Epoch 17/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0119 - val_loss: 0.0053 Epoch 18/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0119 - val_loss: 0.0053 Epoch 19/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0119 - val_loss: 0.0053 Epoch 20/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0118 - val_loss: 0.0053 Epoch 21/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0118 - val_loss: 0.0053 Epoch 22/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0118 - val_loss: 0.0053 Epoch 23/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0118 - val_loss: 0.0053 Epoch 24/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0118 - val_loss: 0.0053 Epoch 25/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0118 - val_loss: 0.0053 Epoch 26/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0118 - val_loss: 0.0053 Epoch 27/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 28/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 29/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 30/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 31/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 32/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 33/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 34/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 35/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 36/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0117 - val_loss: 0.0053 Epoch 37/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0053 Epoch 38/50 79/79 [==============================] - 2s 19ms/step - loss: 0.0116 - val_loss: 0.0053 Epoch 39/50 79/79 [==============================] - 2s 23ms/step - loss: 0.0116 - val_loss: 0.0054 Epoch 40/50 79/79 [==============================] - 2s 20ms/step - loss: 0.0116 - val_loss: 0.0054 Epoch 41/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0054 Epoch 42/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0054 Epoch 43/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0054 Epoch 44/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0054 Epoch 45/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0055 Epoch 46/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0055 Epoch 47/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0055 Epoch 48/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0055 Epoch 49/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0055 Epoch 50/50 79/79 [==============================] - 1s 15ms/step - loss: 0.0116 - val_loss: 0.0055 compilation time : 64.50815987586975
joblib.dump(GRU_model_5p,'/content/drive/Shareddrives/Mineria /GRU_r')
WARNING:absl:Found untraced functions such as gru_cell_2_layer_call_fn, gru_cell_2_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.
['/content/drive/Shareddrives/Mineria /GRU_r']
GRU_model_5p.summary()
Model: "series_GRU_model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_3 (InputLayer) [(None, 50, 1)] 0
GRU (GRU) (None, 60) 11340
dense_2 (Dense) (None, 1) 61
=================================================================
Total params: 11,401
Trainable params: 11,401
Non-trainable params: 0
_________________________________________________________________
GRU_losses_5p = pd.DataFrame(GRU_5p.history)
fig = px.line(GRU_losses_5p, x=GRU_losses_5p.index, y=["loss", "val_loss"])
fig.update_xaxes(title_text="Epoch")
fig.update_yaxes(title_text="Loss")
fig.show()
GRU_Predict_5p = GRU_model.predict(X_test_5p )
GRU_Predict_5p = scaler.inverse_transform(GRU_Predict_5p)
10/10 [==============================] - 0s 6ms/step
# plot y_train, y_test, and testPredict using plotly
seq_len = 50
fig = go.Figure()
fig.add_trace(
go.Scatter(
x=df.index[seq_len : len(y_train_5p) + seq_len],
y=scaler.inverse_transform(y_train_5p).ravel(),
mode="lines",
name="Entrenamiento",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train_5p) + seq_len :],
y=scaler.inverse_transform(y_test_5p).ravel(),
mode="lines",
name="Prueba",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train_5p) + seq_len :],
y=GRU_Predict_5p.ravel(),
mode="lines",
name="PredicciĆ³n",
)
)
fig.update_xaxes(title_text="Fecha")
fig.update_yaxes(title_text="ValorObservado")
fig.show()
GRU_Score2 = metrics.mean_squared_error(scaler.inverse_transform(y_test_5p), GRU_Predict_5p)
print('Test Score: %.2f RMSE' % (GRU_Score2))
Test Score: 1.53 RMSE
def create_dataset(X, y, time_steps=1):
# crea dos listas vacias para depositar los datos
Xs, ys = [], []
# el primer lote de datos empieza en la primera observaciĆ³n
# y toma time_steps datos.
# Comienza a avanzar hacia adelante.
for i in range(len(X) - time_steps):
v = X.iloc[i:(i + time_steps)].values
Xs.append(v)
ys.append(y.iloc[i + time_steps])
return np.array(Xs), np.array(ys)
time_steps1 = 100
# reshape to [samples, time_steps, n_features]
X_train_r, y_train_r = create_dataset(train, train, time_steps1)
X_test_r, y_test_r = create_dataset(test, test, time_steps1)
print("X_train.shape = ", X_train_r.shape)
print("y_train.shape = ", y_train_r.shape)
print("X_test.shape = ", X_test_r.shape)
print("y_test.shape = ", y_test_r.shape)
X_train.shape = (1360, 100, 1) y_train.shape = (1360, 1) X_test.shape = (266, 100, 1) y_test.shape = (266, 1)
# shapes
inputs_shape = (X_train_r.shape[1], X_train_r.shape[2])
GRU_output_r = 60
# layers
inputs_r = L.Input(inputs_shape)
gru_r = L.GRU(units=GRU_output_r, name='GRU')(inputs_r)
outputs_GRU_r = L.Dense(1)(gru_r)
GRU_model_r = Model(inputs=inputs_r, outputs=outputs_GRU_r, name='series_GRU_model')
# Compiling the RNN
GRU_model_r.compile(optimizer=Adam(0.001),loss="mean_squared_error")
# Fitting to the training set
start = time.time()
GRU_r = GRU_model_r.fit(
X_train_r,
y_train_r,
epochs=50,
batch_size=16,
validation_split=0.1,
verbose=1,
shuffle=False
)
print("compilation time : ", time.time() - start)
Epoch 1/50 77/77 [==============================] - 5s 45ms/step - loss: 0.0365 - val_loss: 0.0042 Epoch 2/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0082 - val_loss: 0.0040 Epoch 3/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0073 - val_loss: 0.0039 Epoch 4/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0067 - val_loss: 0.0036 Epoch 5/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0063 - val_loss: 0.0034 Epoch 6/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0061 - val_loss: 0.0033 Epoch 7/50 77/77 [==============================] - 4s 51ms/step - loss: 0.0060 - val_loss: 0.0032 Epoch 8/50 77/77 [==============================] - 4s 47ms/step - loss: 0.0059 - val_loss: 0.0032 Epoch 9/50 77/77 [==============================] - 3s 38ms/step - loss: 0.0059 - val_loss: 0.0031 Epoch 10/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 11/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 12/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 13/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 14/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 15/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 16/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 17/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 18/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 19/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 20/50 77/77 [==============================] - 3s 40ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 21/50 77/77 [==============================] - 4s 49ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 22/50 77/77 [==============================] - 4s 49ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 23/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 24/50 77/77 [==============================] - 4s 58ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 25/50 77/77 [==============================] - 3s 40ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 26/50 77/77 [==============================] - 3s 42ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 27/50 77/77 [==============================] - 4s 57ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 28/50 77/77 [==============================] - 3s 38ms/step - loss: 0.0058 - val_loss: 0.0031 Epoch 29/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 30/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 31/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 32/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 33/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 34/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 35/50 77/77 [==============================] - 5s 59ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 36/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 37/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 38/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 39/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 40/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 41/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 42/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 43/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 44/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 45/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 46/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 47/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 48/50 77/77 [==============================] - 3s 39ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 49/50 77/77 [==============================] - 4s 53ms/step - loss: 0.0057 - val_loss: 0.0031 Epoch 50/50 77/77 [==============================] - 4s 45ms/step - loss: 0.0057 - val_loss: 0.0031 compilation time : 161.8928415775299
joblib.dump(GRU_model_5p,'/content/drive/Shareddrives/Mineria /GRU_r')
WARNING:absl:Found untraced functions such as gru_cell_2_layer_call_fn, gru_cell_2_layer_call_and_return_conditional_losses while saving (showing 2 of 2). These functions will not be directly callable after loading.
['/content/drive/Shareddrives/Mineria /GRU_r']
GRU_model_r.summary()
Model: "series_GRU_model"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
input_4 (InputLayer) [(None, 100, 1)] 0
GRU (GRU) (None, 60) 11340
dense_3 (Dense) (None, 1) 61
=================================================================
Total params: 11,401
Trainable params: 11,401
Non-trainable params: 0
_________________________________________________________________
GRU_losses_r = pd.DataFrame(GRU_r.history)
fig = px.line(GRU_losses_r, x=GRU_losses_r.index, y=["loss", "val_loss"])
fig.update_xaxes(title_text="Epoch")
fig.update_yaxes(title_text="Loss")
fig.show()
GRU_Predict_r = GRU_model_r.predict(X_test_r)
GRU_Predict_r = scaler.inverse_transform(GRU_Predict_r)
9/9 [==============================] - 0s 12ms/step
# plot y_train, y_test, and testPredict using plotly
seq_len1 = 100
fig = go.Figure()
fig.add_trace(
go.Scatter(
x=df.index[seq_len : len(y_train_r) + seq_len1],
y=scaler.inverse_transform(y_train_r).ravel(),
mode="lines",
name="Entrenamiento",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train_r) + seq_len1 :],
y=scaler.inverse_transform(y_test_r).ravel(),
mode="lines",
name="Prueba",
)
)
fig.add_trace(
go.Scatter(
x=df.index[len(y_train_r) + seq_len1 :],
y=GRU_Predict_r.ravel(),
mode="lines",
name="PredicciĆ³n",
)
)
fig.update_xaxes(title_text="Fecha")
fig.update_yaxes(title_text="ValorObservado")
fig.show()
GRU_Score3 = metrics.mean_squared_error(scaler.inverse_transform(y_test_r), GRU_Predict_r)
print('Test Score: %.2f RMSE' % (GRU_Score3))
Test Score: 0.65 RMSE
print('Modelo GRU 1 paso adelante 50 retardos %f' % (GRU_Score1))
print('Modelo GRU 5 paso adelante 50 retardos %f' % (GRU_Score2))
print('Modelo GRU 1 paso adelante 100 retardos %f' % (GRU_Score3))
Modelo GRU 1 paso adelante 50 retardos 0.835264 Modelo GRU 5 paso adelante 50 retardos 1.525437 Modelo GRU 1 paso adelante 100 retardos 0.649107